Aaron Stone wrote: > I probably broke it. Take a look and tell me if you think passing the > "mailbox source" will end up making sense in the context of username > [EMAIL PROTECTED] addressing, which is directly supported by Sieve's > address tests and so now is a good time to make it work under this > algorithm: > > - User-specified (e.g. by command line option to dbmail-smtp or > in a Sieve script) non-existant mailboxes are always created. > - Non-specified mailboxes go to INBOX, which is created. > - Externally specified mailboxes which go to an existing mailbox > are respected, while non-existant mailboxes go to INBOX (and > the INBOX creation rules are followed).
I really like your mailbox_source_t idea. It solves several problems I had with David Nibletta's address+mailbox patch. In addition to INBOX, David's idea to allow auto-creation of runtime specified mailboxes like Spam also deserves condideration. > > Aaron > > > On Sat, 2005-11-26 at 01:41 +0100, Paul J Stevens wrote: > >>Aaron, >> >>I noticed messages are not inserted when the insertion requires mailbox >>creation >>first. Could this be something you did? Else I'll look into it myself. >> >> > > > _______________________________________________ > Dbmail-dev mailing list > Dbmail-dev@dbmail.org > http://twister.fastxs.net/mailman/listinfo/dbmail-dev > -- ________________________________________________________________ Paul Stevens mailto:[EMAIL PROTECTED] NET FACILITIES GROUP PGP: finger [EMAIL PROTECTED] The Netherlands________________________________http://www.nfg.nl
